Draper and Smith (1966, p. 41) considered data on the relationship between Table 12.18: Tetrahydrocortisone values for patients with Cushing’s syndrome.

a b c 3.1 8.3 15.4 10.2 3.0 3.8 7.7 9.2 1.9 3.9 6.5 9.6 3.8 7.8 5.7 53.8 4.1 9.1 13.6 15.8 1.9 Table 12.19: Age and costs of maintenance for truck tractors.
Age Costs 0.5 163 182 1.0 978 466 549 4.0 495 723 681 4.5 619 1049 1033 5.0 890 1522 1194 5.5 987 6.0 764 1373 the age of truck tractors (in years) and the cost (in dollars) of maintaining them over a six-month period. The data are given in Table 12.19.
Note that there is only one observation at 5.5 years of age. This group does not yield an estimate of the variance and can be ignored for the purpose of computing the mean squared error. In the weighted average of variance estimates, the variance of this group is undefined but the variance gets 0 weight, so there is no problem.
Give the analysis of variance table for these data. Does cost differ with age? Is there a significant difference between the cost at 0.5 years as opposed to 1.0 year? Determine whether there are any differences between costs at 4, 4.5, 5, 5.5, and 6 years. Are there differences between the first two ages and the last five? How well do polynomials fit the data?